Gender convergence in core housework hours

Abstract

The increase in dual-earner families over the last few decades raises questions about whether men are sharing the homemaker role now that women are sharing the breadwinner role. Theories of the allocation of unpaid work, such as dependency and exchange approaches, time availability and gender display, were developed during the period when the male-breadwinner family was the dominant family type in many Western societies. Using data collected by three Australian surveys conducted between 1986 and 2005 (N = 5598), I find that the increase in dual-earner families has been accompanied by a convergence in housework hours due to men spending more time and women spending less time doing core housework tasks. Attitudes to gender roles appear to be the key predictor of housework hours overriding the effects of time availability, income and relative income

Keywords

data analysis family gender gender roles quantitative analysis sociology of the family

During the latter half of the 20th century, the typical Australian family moved from being a male-breadwinner family to a dual-earner family due to a number of social and economic changes, including the restructuring of the labour market, women’s increased levels of human capital, the removal of barriers preventing women’s employment and recognition of women’s entitlement to equal pay for equal work. By the end of the century, an increase in the labour force participation rate of married women meant that the majority of families were providing two workers to the labour market. To cope with the demands of family and the need to earn an income, the majority of Australian mothers take up part-time employment.

According to the Australian Bureau of Statistics (ABS), in 2007 38 percent of Australian couple families with dependent children aged 0–14 years were dual-earner families in which the mother was employed part-time and a further 23 percent were dual-earner families in which the mother was employed full-time (ABS, 2008). The trend towards dual-earner families has renewed interest in the allocation of housework within families, leading some researchers to examine whether the sharing of the breadwinner role, by providing income, has been accompanied by a sharing of the homemaker role, providing the unpaid labour essential to keep families functioning. The major theoretical perspectives used to explain the allocation of housework within couple families were developed during the era when the male-breadwinner family was the dominant family type. Now that dual-earner families outnumber male breadwinners, it is timely to consider the relevance of these theories.

Theoretical perspectives

The time availability theory suggests that the partner with the most ‘free’ time will make the rational decision to assume responsibility for housework tasks (Blood and Wolfe, 1960). In single-earner families, one partner works full-time and therefore has less time available for housework than the other partner, who has no paid work commitments and can devote their time to housework. Although there is evidence that, for both men and women, time spent in paid employment has a negative effect on housework hours (Baxter et al., 2005; Bittman et al., 2003; Ferree, 1991), several researchers have found that women continue to do the bulk of housework regardless of their share of the couple’s combined paid work hours (Baxter, 2002; Baxter et al., 2005; Bond and Sales, 2001; Coltrane and Ishii-Kuntz, 1992; Sullivan, 2006). There is also evidence that women in paid employment do more housework than their unemployed male partners and that unemployed men in couple families do less housework than employed men in couple families (Brines, 1994).

In A Treatise on the Family, Becker (1981) argued that the division of labour in the male-breadwinner/female-homemaker family is a rational choice based on sound economic principles. According to Becker, men have a comparative advantage in market labour due to their higher levels of human capital and women have a comparative advantage in domestic labour due to their biology. The gender division of labour in the family is therefore the most efficient and rational way to utilize the resources of both partners and maximize family outcomes. From this perspective, the allocation of work is based on overall household efficiency, assuming that it is more efficient from a household perspective if the male partner concentrates his efforts on market work while the female partner concentrates her efforts on household work. Folbre (2004: 7) argues that Becker’s theory is flawed, overstating the role of individual decisions and exaggerating the efficiency of social outcomes. Becker downplays the constraining role of institutions that structure the context within which individuals make their decisions. For example, in Australia, many of the elements of men’s comparative advantage in the labour market were the result of institutional barriers to women’s employment, including the ‘family wage’¹ for men and not women and the ‘marriage bar’² legislation limiting married women’s access to paid employment.

The exchange of economic resources theory is also based on the assumption that partners in couple families make rational decisions regarding the division of labour (Baxter, 2002; Coltrane, 2000). The female partner exchanges her contribution of unpaid domestic labour for the economic resources provided by the male partner (Brines, 1994). There is evidence, however, that the exchange of economic resources theory cannot explain the allocation of domestic labour within dual-earner or female-breadwinner families; family types that have become increasingly common during the past few decades. For example, Brines (1994: 682) found that when husbands were dependent on the income of their wives, they were less willing to perform unpaid labour in exchange for the economic resources provided by their wives. Artis and Pavalko (2003) found that the relative resources of the wife had no effect on the allocation of housework, concluding that, net of life course and time availability factors, an increase in the relative proportion of household income provided by the wife did not equate to a significant change in her responsibility for housework.

Furthermore, Bittman et al. (2003) found that female partners decreased their housework hours only up to the point where they contributed an equal share of the household income. Once they contributed the majority of the household income, their share of housework increased. Bittman et al. (2003) concluded that these women compensated for their deviance from accepted gender norms (that is, the need to be dependent on the income of their husbands) by doing extra housework. According to Bittman et al. (2003: 210) ‘gender trumps money’, that is, the effect of gender expectations on the division of housework was powerful enough to reverse the effects of relative resources (see also Baxter et al., 2005; Brines, 1994; Greenstein, 2000; Hook, 2006). Hochschild (1989: 221) argued that when the wife earned more than the husband, his identity was ‘financially threatened’, therefore he would do less housework in order to protect his masculinity. So it would seem that the effect of relative resources is contingent upon how far the couple’s relative earnings stray from the norm.

According to Gupta (2007), the inability of the exchange of economic resources measures to capture the association between relative resources and housework hours was due to a flaw in the thinking underpinning the concept of relative resources. From Gupta’s perspective, the effect of a unit change in the income of the male partner was not necessarily equal to the effect of a unit change in the female partner’s income. Gupta’s analyses of National Survey of Families and Households (NSFH) data showed that women’s relative earnings explained little of the variance in housework hours when women’s absolute earnings were factored in. He concluded that women’s housework time depended only on their own earnings rather than the ‘magnitude of their earnings compared to their husband’s’ (2007: 412).

From an alternative perspective, the allocation of paid and unpaid work within couple families is linked to West and Zimmerman’s (1987: 146) concept of ‘doing gender’. West and Zimmerman argue that ‘if we do gender appropriately, we simultaneously sustain, reproduce, and render legitimate the institutional arrangements that are based on sex category’. That is, by behaving in the particular ways that society accepts as indicative of masculine behaviour one is ‘doing gender’ in a way that will announce that one is a man and by behaving in the particular ways that society accepts as indicative of feminine behaviour one is ‘doing gender’ in a way that will announce that one is a woman. Doing specific household tasks provides opportunities to demonstrate to others that one is a competent member of a particular sex category (Coltrane, 2000). That is, when males take care of the car and home repairs, they are performing masculinity and when females take care of the cooking and cleaning, they are performing femininity.

Some researchers link the performance or avoidance of housework to one’s gender ideology. Gender ideologies are linked to beliefs about appropriate behaviour for men and women, particularly with regard to marital roles (Morehead, 2005). According to Greenstein (2000: 323) gender identity is one’s self-definition as being either a man or a woman whereas gender ideologies are the ‘elements that make up that definition’. That is, whether one believes men should do certain tasks and women should do other tasks is independent of one’s own identity as either a man or a woman. For example, for some men, housework is ‘women’s work’ that should be left for mothers, wives or daughters to take care of, while for other men, housework is the responsibility of every person within the household. Individuals who subscribe to the ‘housework is women’s work’ perspective are regarded as holding traditional gender ideologies whereas those who see housework as each individual’s responsibility are regarded as holding egalitarian gender ideologies.

The gender display perspective predicts that men with more traditional gender attitudes will spend less time doing housework than men with more egalitarian gender attitudes, and women with more traditional gender attitudes will spend more time doing housework than women with more egalitarian gender attitudes. Several researchers, both here in Australia and overseas have found evidence of this (Artis and Pavalko, 2003; Baxter, 2002; Baxter et al., 2005; Bianchi et al., 2000; Blair and Lichter, 1991; Casper and Bianchi, 2002; Cunningham, 2005; Gershuny et al., 1994; Greenstein, 2000). Casper and Bianchi (2002) suggest that this is because couples with egalitarian gender attitudes regard housework tasks as the responsibility of both partners and are therefore more likely to enjoy a more egalitarian division of domestic labour.

In sum, theories based on the exchange of economic resources were able to explain the allocation of housework tasks within couple families when the male-breadwinner family was the modal family type and gender display theories can provide some insight into why exchange of economic resources theories are no longer reliable now that the majority of couple families are dual-earner families. In this article, I seek to answer two research questions: Has the trend towards dual-earner families had any impact on the core housework hours of men and women in couple families in Australia? And do measures of the theories commonly used to explain the allocation of housework within couple families remain relevant now that the majority of couple families are dual-earner families.

Method

Data

Data for this study come from three nationally representative cross-sectional surveys: the 1986 Social Structure of Australia Project, the 1993 Social Structure of Australia Project and the 2005 Neoliberalism, Inequality and Politics Project. The 1986 survey was restricted to male respondents working at least 30 hours per week and female respondents working at least 15 hours per week (N = 1195). All respondents were aged 18 years or over and the data were collected via face-to-face interviews and a self-complete questionnaire (J.S. Western et al., 2005). The 1993 survey collected data via a self-complete mail-out questionnaire from 2780 respondents aged 18 years or over (J.S. Western et al., 1993). The 2005 survey collected data from a national random sample of 1623 individuals aged 18 years or over via computer-assisted telephone interviews (M. Western et al., 2005). In each survey, only one adult per household was interviewed.

For reasons of comparability across the various project samples, my analytical sample is restricted to male respondents employed 30 or more hours per week and female respondents employed 15 or more hours per week. Although this limits the generalizability of the findings, using the 1986 sample parameters allows me to compare changes over 19 years (from 1986 to 2005) rather than 12 years (from 1993 to 2005). These three data sets are the only ones available in Australia that provide data on time spent doing housework, gender attitudes and the demographic characteristics usually included in the analysis of housework hours.

Dependent variable

Each of the three surveys collected data on the amount of time respondents reported spending on various housework tasks. By summing the hours and minutes spent doing six housework tasks: preparing meals, doing the dishes, grocery shopping, cleaning the house, washing and ironing, I constructed a core housework index. Although this list is not exhaustive, it does include the commonly measured core housework tasks which account for the bulk of all housework and which need to be completed on a regular, sometimes even daily basis (Coltrane, 2000). The data analysed for this study also include measures of four other housework tasks: taking out the garbage, mowing the lawn, gardening and home maintenance and repairs. These tasks are not generally regarded as core tasks as they are often carried out on a less frequent, basis, if at all. For example, a couple living in an apartment does not need to spend time mowing the lawn. On the other hand, for many people gardening is a form of leisure, a task they spend extra time doing for pleasure. Including these extra four tasks makes no difference to the substantive results³ so I restrict my analyses to the core tasks in line with those of other researchers (Bond and Sales, 2001; Coltrane and Ishii-Kuntz, 1992; Sullivan, 2000).

Predictor variables

To determine the salience of three theoretical perspectives: time availability, exchange of economic resources and gender display, I include four predictor variables: family type, relative income, absolute income and gender attitudes. Family type is used as a measure of the time availability of both partners and is included as a series of dummy variables: co-breadwinner (both partners employed full-time) (reference category); new traditional (male partner employed full-time and female partner employed part-time); male breadwinner (male partner employed full-time and female partner not in paid employment); and other (male partner employed part-time or not employed). Due to the restriction of this sample to employed men and women, there are no female respondents in male-breadwinner families. Respondents are grouped according to whether they were employed part-time (15–34 hours) or full-time (35 hours or more).⁴ Partners are divided into three categories according to their hours of employment: no paid employment; part-time paid employment (1–34 hours); and full-time paid employment (35 hours or more).

The exchange of economic resources theory suggests that relative income is an important indicator of housework hours, with the partner providing the greater share of total income doing less housework. To assess whether respondents providing a greater proportion of their household’s total income performed less housework than respondents providing a smaller proportion of their household’s total income, I created a measure of the proportion of income contributed by the respondent to the household by dividing the respondent’s income by the total household income. I also include a dummy variable, coded 1 = missing, on income proportion in the regression analyses to identify respondents who did not provide data on their income and/or their household income.

Although most previous research on housework hours has focused on the relative contribution to total household income provided by the respondent, Gupta (2006, 2007) argued that the respondents’ absolute income should also be considered. There are 130 respondents with missing data on income. Although this level of missing data is unfortunate, it is not unusual for respondents to decline to give information about their incomes. Rather than drop these respondents from the analyses, I calculated imputed values for respondents’ income using the ‘impute’ command in Stata (2007) and adding sex, age, occupation and employment status as independent variables. The impute command organizes the cases by patterns of missing data so that missing values are imputed using the prediction from the best available subset of the data (Stata, 2007: 268–9). To overcome the skewed distribution of income, I take the log of income.

Gender display theories suggest that gender attitudes are an important indicator of housework hours. In previous studies, researchers have typically found that respondents with more egalitarian gender attitudes enjoy a more egalitarian division of domestic labour (Baxter, 2002; Baxter et al., 2005; Bianchi et al., 2000; Coltrane, 2000; Coltrane and Ishii-Kuntz, 1992; Gershuny et al., 1994; Greenstein, 2000). Gender attitudes are measured using an index based on four questions that were included in all three surveys. Respondents were asked to indicate on a scale of 1 to 4 how strongly they agreed or disagreed with the following statements:

It is better for the family if the husband is the principal breadwinner outside the home and the wife has primary responsibility for the home and children.

If both the husband and wife work, they should share equally in the housework and childcare.

Ideally, there should be as many women as men in important positions in government and business.

There should be satisfactory childcare facilities so that women can take jobs outside the home.

Responses to question one were reverse coded so that low scores on all items indicated a more egalitarian gender ideology. The gender attitudes index is constructed by summing the responses to the four questions and then dividing the total by the number of valid responses and ranges from 1 (egalitarian attitudes) to 4 (traditional attitudes). The internal reliability of this index was tested and returned a Cronbach’s alpha of 0.68 in 1986, 0.62 in 1993 and 0.64 in 2005. These results indicate that the four questions are all measuring the same underlying concept, therefore providing evidence of the overall internal coherence of these items. The Cronbach’s alpha scores do not fluctuate greatly over time, indicating the external reliability of this index. I include a dummy variable, coded 1 = missing data, on the gender attitudes index in the regression analysis.

Control variables

Previous researchers have identified several background characteristics that impact on housework hours, including: age, education, presence of dependent children and home ownership. Measures of these attributes are included as controls. Age is included as a series of dummy variables: 18–34 years (reference category); 35–44 years; 45–54 years; 55–64 years. The dependent child variable is coded 1 = child aged under 15 years living in the household. Education is included as a series of dummy variables differentiating respondents according to their highest level of education: less than year 12 education (reference category); completed year 12; trade certificate/diploma; university degree. Home ownership is coded 1 = renting or other. Dummy variables identifying missing values on each of the control variables are also included in the analyses. Table A1 in the Appendix lists the descriptive statistics for each of the variables of interest.

Analytical strategy

The analyses are conducted in two phases. First, to answer the first research question, I examine whether there has been any convergence in the core housework hours of employed men and women in couple families. Using Stata, I include income, relative income and gender attitudes as well as each of the control variables to estimate the predicted means of housework hours for men and women in each family type in each year. In the second phase of the analysis, I conduct ordinary least squares regression analysis using a series of models to determine the relevance of various measures predicted by the literature to affect the housework hours of employed men and women in couple families. The first model includes family type and the four control variables. In the second model, I replace family type with relative income. In the third model, I replace relative income with gender attitudes. The final model includes all of the predictor and control variables to simultaneously test the salience of the various theories commonly used to explain the gender gap in housework hours. Previous research has found that gender accounts for a large proportion of the variance in housework hours and that the predictor and control variables selected for this study have different effects on the housework hours of men and women (Baxter et al., 2005; Bianchi et al., 2000; Davis and Greenstein, 2004), therefore, I include interaction terms for sex and each of the predictor and control variables.

Results

The predicted means for men’s and women’s housework hours in each family type in each year are presented in Table A2 in the Appendix. The graph of the predicted means is presented as Figure 1. In each year, men in ‘other’ families, that is, men not in full-time paid employment spent more time doing housework compared to other men and women in co-breadwinner families spent less time doing housework than other women. Men in male-breadwinner families spent the least amount of time doing housework and women in new traditional families spent the most time doing housework. Housework hours of men in all family types increased steadily between 1986 and 2005. For example, men in new traditional families averaged 8.4 hours per week in 1986, 9.1 hours per week in 1993 and 10 hours per week in 2005. On the other hand, women’s time spent doing housework increased between 1986 and 1993 before declining between 1993 and 2005. For example, women in new traditional families averaged 32.7 hours per week in 1986, 35.3 hours per week in 1993 and 27.7 hours per week in 2005. Overall convergence in time spent doing housework is due to an increase in men’s time and a decrease in women’s time.

Figure 1:

Predicted values for core housework, controlling for the effects of all predictor and control variables and gender interactions

Turning now to the salience of the various theories commonly used to explain the core housework hours of men and women in couple families, I present the results of a series of linear regression models. Although interaction terms were created for sex with each of the predictor and control variables, only interaction terms which produced statistically significant coefficients are included in the models presented here. The first model tests the association between time availability and time spent doing housework and the results are shown in Table 1. As predicted by the literature, women spent more time doing housework than men in each year. Men in male-breadwinner families spent less time doing housework than men in co-breadwinner families in each year. Women in new traditional families spent more time doing housework than women in co-breadwinner families. Women with a university degree spent less time doing housework than other women in each year. The presence of dependent children is associated with an increase in housework hours in 1993 and 2005 for both men and women and for women in 1986.

Table 1:

Model 1 testing the relevance of time availability

	1986		1993		2005
Core housework	Coeff.	Std err.	Coeff.	Std err.	Coeff.	Std err.
Male (ref.)
Female	13.73***	2.03	15.86***	2.02	10.18**	3.71
Co-breadwinner (ref.)
New traditional	−1.84	1.51	−3.23*	1.27	−1.97	1.56
Male breadwinner	−3.08*	1.25	−3.91***	1.16	−5.55***	1.59
Other	−0.32	4.09	−1.23	3.43	5.21	3.51
18–34 years (ref.)
35–44 years	−1.91	1.19	−0.07	1.21	−1.46	2.00
45–54 years	−2.21	1.40	−2.00	1.35	−1.10	1.96
55–64 years	−2.75	1.95	−2.01	1.71	0.21	2.19
Dependent child 1 = yes	−0.41	1.19	2.16*	1.08	3.05*	1.47
<Year 12 (ref.)
Year 12	0.57	1.32	−0.38	1.46	1.61	2.48
Certificate/diploma	−1.06	1.38	−2.31*	1.17	1.24	2.28
University degree	−0.32	1.30	−0.06	1.43	1.42	2.21
Buy home (ref.)
Other	0.24	0.94	0.75	1.08	−0.26	1.36
Sex interactions
Female × new traditional	9.71***	2.01	12.08***	1.84	7.02***	2.18
Female × other	4.20	4.51	3.94	3.67	−0.50	3.92
Female × 35–44 years	3.01	1.79	0.15	1.83	4.73	2.70
Female × 45–54 years	8.19***	2.25	3.35	2.00	1.94	2.64
Female × 55–64 years	6.42*	3.25	6.41*	2.85	−0.75	3.07
Female × dependent child	6.90***	1.78	3.55*	1.71	2.47	2.12
Female × Year 12	−8.45***	2.01	−6.95**	2.23	−3.43	3.44
Female × Certificate/diploma	−4.05	2.25	−2.46	1.87	−4.08	3.22
Female × university degree	−6.19**	2.05	−11.81***	2.10	−7.35*	3.15
Constant	11.58***	1.56	11.73***	1.48	10.46***	2.76
Adj. R-squared	0.5117		0.5380		0.3392
N	813		1056		585

The coefficients for ‘missing’ on each of the control variables were calculated but have not been included in this table.

The salience of the exchange of economic resources theory is tested in Model 2. As Table 2 shows, relative income is negatively associated with time spent doing housework in 1986 and 2005 for both men and women, and for women but not men in 1993. Being female is positively associated with housework hours in 1986 and 1993, net of other factors. There is also an effect for age for women, with older women spending more time doing housework in 1986 and 1993. University-educated women spent less time doing housework in each year and the presence of dependent children increased the time spent doing housework for women in 1986 and 1993 and for both men and women in 2005. Replacing relative income with log of absolute income had no effect (results not shown).

Table 2:

Model 2 testing the relevance of the exchange of economic resources

	1986		1993		2005
Core housework	Coeff.	Std err.	Coeff.	Std err.	Coeff.	Std err.
Male (ref.)
Female	15.41***	3.15	21.86***	3.02	8.70	4.50
Relative income	−0.07**	0.03	−0.04	0.03	−0.11***	0.03
18–34 years (ref.)
35–44 years	−1.75	1.21	−0.16	1.22	−1.83	2.02
45–54 years	−2.13	1.42	−2.48	1.35	−1.23	1.97
55–64 years	−2.60	1.98	−3.05	1.69	−0.27	2.17
Dependent child 1 = yes	−0.37	1.18	1.16	1.05	2.92*	1.48
<Year 12 (ref.)
Year 12	0.66	1.34	−0.23	1.47	2.28	2.49
Certificate/diploma	−0.93	1.40	−2.10*	1.18	1.78	2.28
University degree	−0.13	1.32	0.41	1.45	2.13	2.20
Buy home (ref.)
Not own home	0.49	0.97	0.94	1.09	0.61	1.37
Sex Interactions
Female × relative income	0.0003	0.04	−0.08*	0.04	0.04	0.04
Female × 35–44 years	3.07	1.83	0.10	1.85	5.88*	2.71
Female × 45–54 years	9.01***	2.30	4.49*	2.00	3.23	2.63
Female × 55–64 years	8.49*	3.31	7.62**	2.84	2.38	2.99
Female × dependent child	8.20***	1.78	6.27**	1.65	3.63	2.10
Female × year 12	−8.82***	2.03	−7.04**	2.26	−3.72	3.46
Female × Certificate/diploma	−4.56*	2.28	−1.25	1.90	−4.01	3.24
Female × university degree	−6.73***	2.08	−10.95***	2.16	−7.10*	3.17
Constant	15.14***	2.48	13.49***	2.33	15.79***	3.44
Adj. R-squared	0.4961		0.5282		0.3270
N	813		1056		585

The coefficients for ‘missing’ on each of the control variables were calculated but have not been included in this table.

The results from Model 3 testing the association between gender attitudes and housework hours for men and women in each year are shown in Table 3. Gender attitudes are negatively associated with the housework hours of men and positively associated with the housework hours of women in each year. Women with dependent children reported spending more time doing housework in each year, as did older women in 1986 and 1993. Women with a university degree spent less time doing housework.

Table 3:

Model 3 testing the relevance of gender attitudes

	1986		1993		2005
Core housework	Coeff.	Std err.	Coeff.	Std err.	Coeff.	Std err.
Male (ref.)
Female	7.96*	3.71	9.48**	3.66	−0.43	4.76
Gender attitudes	−3.52***	0.99	−2.74**	1.00	−3.64**	1.24
18–34 years (ref.)
35–44 years	−0.99	1.22	0.04	1.24	−0.96	2.01
45–54 years	−1.15	1.46	−1.78	1.40	−0.09	1.98
55–64 years	−1.77	2.00	−2.51	1.72	0.12	2.18
Dependent child 1 = yes	−1.00	1.14	1.20	1.05	1.72	1.43
<Year 12 (ref.)
Year 12	0.40	1.35	−0.19	1.49	2.32	2.48
Certificate/diploma	−0.24	1.41	−2.27	1.19	2.03	2.27
University degree	−0.05	1.33	0.06	1.46	1.92	2.20
Buy home (ref.)
Not own home	0.09	0.95	0.92	1.10	−0.29	1.37
Sex Interactions
Female × gender attitudes	4.81**	1.65	5.07**	1.61	7.99***	1.80
Female × 35–44 years	2.24	1.83	0.16	1.88	4.09	2.72
Female × 45–54 years	8.04***	2.33	4.86*	2.03	1.13	2.67
Female × 55–64 years	6.67*	3.34	7.32*	2.87	−0.28	3.07
Female × dependent child	9.37***	1.74	7.48***	1.65	4.82*	2.05
Female × year 12	−8.41***	2.04	−7.98***	2.27	−3.86	3.45
Female × Certificate/diploma	−4.89*	2.29	−2.30	1.91	−4.12	3.23
Female × university degree	−6.79***	2.11	−11.93***	2.17	−6.89*	3.16
Constant	16.97***	2.52	15.41***	2.44	15.18***	3.57
Adj. R-squared	0.4968		0.5184		0.3307
N	813		1056		585

The coefficients for ‘missing’ on each of the control variables were calculated but have not been included in this table.

In the final model all of the predictor and control variables are included to test the salience of the three theoretical perspectives simultaneously. The results are shown in Table 4. The most consistent predictor of housework hours is gender attitudes. Men with more traditional attitudes spent less time doing housework and women with more traditional attitudes spent more time doing housework than their more egalitarian counterparts. Relative income and absolute income had no effect for either men or women in any year. Although the family type variables did produce some statistically significant coefficients, the pattern is not consistent across the three datasets.

Table 4:

Model 4 testing the relevance of all four theories

	1986		1993		2005
Core housework	Coeff.	Std err.	Coeff.	Std err.	Coeff.	Std err.
Male (ref.)
Female	25.90	17.85	29.01*	12.02	9.16	19.99
Co-breadwinner (ref.)
New traditional	−0.01	1.68	−3.03*	1.27	−0.60	1.62
Male breadwinner	1.06	2.49	−3.17*	1.28	−3.79	2.04
Other	2.79	4.27	−1.32	3.50	4.91	3.49
Relative income	−0.09	0.05	0.001	0.03	−0.03	0.04
Log of income	1.24	1.21	−0.81	0.86	−1.95	1.45
Gender attitudes	−3.33***	0.97	−2.28*	0.99	−3.02*	1.24
18–34 years (ref.)
35–44 years	−1.24	1.23	0.27	1.20	−1.42	1.98
45–54 years	−1.12	1.43	−1.27	1.37	−0.71	1.95
55–64 years	−1.32	1.98	−1.60	1.71	0.71	2.16
Dependent child 1 = yes	−0.14	1.20	2.03	1.08	3.10*	1.46
<Year 12 (ref.)
Year 12	0.21	1.33	−0.36	1.46	1.59	2.43
Certificate/diploma	−0.44	1.39	−2.19	1.17	1.83	2.25
University degree	−0.73	1.41	−0.01	1.52	1.83	2.24
Buy home (ref.)
Other	0.15	0.95	0.60	1.07	−1.04	1.37
Sex Interactions
Female × new traditional	6.27**	2.30	9.57***	1.94	3.04	2.33
Female × other	4.47	4.97	2.71	3.74	−1.06	3.97
Female × relative income	0.01	0.07	−0.03	0.04	0.001	0.06
Female × income	−1.95	1.66	−1.87	1.21	−0.92	1.92
Female × gender attitudes	3.77*	1.63	3.36*	1.58	6.23***	1.80
Female × 35–44 years	2.17	1.84	0.05	1.83	4.19	2.69
Female × 45–54 years	6.56**	2.31	2.51	2.02	0.63	2.65
Female × 55–64 years	5.00	3.33	6.22*	2.85	−2.49	3.10
Female × dependent child	6.11***	1.82	3.21	1.70	1.54	2.11
Female × year 12	−8.23***	2.02	−6.37**	2.23	−4.14	3.39
Female × Certificate/diploma	−4.53*	2.26	−1.06	1.89	−3.84	3.19
Female × university degree	−5.20*	2.18	−9.02***	2.26	−5.75	3.19
Constant	9.52	12.98	24.64**	8.60	38.53*	15.09
Adj. R-squared	0.5196		0.5481		0.3630
N	813		1056		585

The coefficients for ‘missing’ on each of the control variables were calculated but have not been included in this table.

Discussion

This study set out to determine whether the trend towards dual-earner families had been accompanied by a convergence in housework hours and whether the various theories commonly used to explain the allocation of housework within couple families remained relevant. For men, housework hours are associated with the paid work hours of their partners, whereas for women, housework hours are associated with their own paid work hours. There is also evidence that these associations have become more established over time. For example, men in male-breadwinner families spent 5.5 hours less per week doing housework than men in co-breadwinner families in 2005, whereas in 1986 the difference was 3 hours. However, the effect of family type becomes insignificant once income, relative income and gender attitudes are included in the model, indicating that the relationship between time availability and housework is mediated by the effects of income and gender attitudes.

Relative income does not provide a consistent pattern of effect, however, gender attitudes has the predicted effect for both men and women in each of the three years. Although the effect declined slightly between 1986 and 2005 for men, those with more traditional attitudes spent less time doing housework than other men. Women with more traditional attitudes spent more time doing housework than other women and this effect increased over time. When all four predictor variables were included in the model, gender attitudes overrides the effect of relative income, income and time availability. These results suggest that the ability of theories based on the exchange of economic resources to explain the allocation of housework have been overstated. Consequently, measures of gender attitudes would seem to be a necessary inclusion in models predicting housework hours. There is also evidence that the variables included in these models explained less of the variation in housework hours in 2005 than they did in 1986. In 1986 these measures explained approximately half of the variation in housework hours, however by 2005 the same measures explained around one-third of the variation.

Overall this article does provide evidence that the trend towards dual-earner families has been accompanied by a decline in women’s and an increase in men’s housework hours, suggesting that as women have become increasingly likely to share the breadwinner role, men have become increasingly likely to share the homemaker role. These results contradict previous Australian research conducted by Baxter (2002) and Bittman and Pixley (1997) suggesting that the trend towards convergence in housework hours due to a decline in the housework hours of women and an increase in the housework hours of men identified in international studies (Bianchi et al., 2000, 2006; Sayer, 2005; Sullivan, 2000) is now being reflected in Australian households. Between 1965 and 1995, housework hours in the US converged as American women reduced their housework time by 12.5 hours per week and American men increased their housework time by 5 hours per week (Bianchi et al., 2000). Housework hours also converged in Canada, the Netherlands, Norway, the UK, the US and Finland between 1961 and 1995, as women decreased and men increased their time spent doing the cooking, cleaning and laundry (Sullivan, 2000). International studies have data dating from the 1960s; in Australia, however, data for the period prior to 1986 is not available and perhaps it is only now that data is available for a 19-year period that I have been able to detect trends in the housework hours of Australian men and women that are similar to those identified in international research.

Conclusion

The results presented here suggest that there has been a general trend for men to spend more time and for women to spend less time doing core housework. Although time availability continues to have some effect on the housework hours of men and women in couple families, men in all family types spent more time and women in all family types spent less time doing housework in 2005 than their counterparts in 1986. The finding that gender attitudes is the only predictor variable to consistently produce statistically significant coefficients in the final model over time suggests that one’s attitude regarding who is responsible for doing the housework is more important than one’s income, relative income or time availability. The decline over time in the amount of variation explained by these measures indicates that a large proportion of the convergence of housework hours identified here is explained by factors not included in these analyses and therefore further investigation is warranted.

Footnotes

Appendix

Table A2:

Predicted values for core housework controlling for the effects of all predictor and control variables and gender interactions.

	1986		1993		2005
Family type	Men	Women	Men	Women	Men	Women
Co-breadwinner	10.48	23.32	11.16	25.14	12.01	18.07
New traditional	8.44	32.72	9.12	35.30	9.98	27.65
Male breadwinner	7.67		8.35		9.21
Other	11.78	25.48	12.46	29.48	13.32	22.57

Funding and Acknowledgements

The Neoliberalism, Inequality and Politics Project from which this article is derived comes from the University of Queensland Social Research Centre and was supported under the Australian Research Council’s Discovery Projects funding scheme (DP0449516). I would like to thank Janeen Baxter and Mark Western for their comments on earlier drafts of this article.

Notes

Biographical note

Jenny Chesters is a research fellow at the University of Canberra. Her current interests include gender inequalities within families and the labour market, the persistence of inequality in educational outcomes and global wealth inequality. Address: Faculty of Education Room 5B3, Building 5, University of Canberra, ACT 2601 Australia. [email: jenny.chesters@canberra.edu.au]

References

Artis

Pavalko

(2003) ‘Explaining the Decline in Women’s Household Labour: Individual Change and Cohort Differences’, Journal of Marriage and Family 65(3): 746–61.

ABS (Australian Bureau of Statistics) (2008) Family Characteristics and Transitions 2006–07 Cat. No. 4442.0. URL (consulted January 2009): www.abs.gov.au

Baxter

(2002) ‘Patterns of Change and Stability in the Gender Division of Household Labour in Australia, 1986–1997’, Journal of Sociology 38(4): 399–424.

Baxter

Hewitt

Western

(2005) ‘Post-familial Families and the Domestic Division of Labour’, Journal of Comparative Family Studies 36(4): 583–600.

Becker

(1981) A Treatise on the Family. Cambridge, MA: Harvard University Press.

Bianchi

S.M.

Robinson

J.P.

Milkie

M.A.

(2006) The Changing Rhythms of American Life. New York: Russell Sage Foundation.

Bianchi

S.M.

Milkie

M.A.

Sayer

L.C.

Robinson

J.P.

(2000) ‘Is Anyone Doing the Housework? Trends in the Gender Division of Household Labor’, Social Forces 79(1): 191–228.

Bittman

Pixley

(1997) The Double Life of the Family: Myth, Hope and Experience. St Leonards: Allen and Unwin.

Bittman

England

Sayer

Folbre

Matheson

(2003) ‘When Does Gender Trump Money? Bargaining and Time in Household Work’, American Journal of Sociology 109(1): 186–214.

10.

Blair

S.L.

Lichter

D.T.

(1991) ‘Measuring the Division of Household Labor: Gender Segregation of Housework Among American Couples’, Journal of Family Issues 12(1): 91–113.

11.

Blood

R.O.

Jr Wolfe

D.M.

(1960) Husbands and Wives: The Dynamics of Married Living. New York: The Free Press.

12.

Bond

Sales

(2001) ‘Household Work in the UK: An Analysis of the British Household Panel Survey 1994’, Work, Employment and Society 15(2): 233–50.

13.

Brines

(1994) ‘Economic Dependency, Gender and the Division of Labor at Home’, American Journal of Sociology 100(3): 652–88.

14.

Casper

L.M.

Bianchi

S.M.

(2002) Continuity and Change in the American Family. Thousand Oaks, CA: Sage.

15.

Coltrane

(2000) ‘Research on Household Labor: Modelling and Measuring the Social Embededness of Routine Family Work’, Journal of Marriage and the Family 62: 1208–33.

16.

Coltrane

Ishii-Kuntz

(1992) ‘Men’s Housework: A Life Course Perspective’, Journal of Marriage and the Family 54(1): 43–57.

17.

Cunningham

(2005) ‘Gender in Cohabitation and Marriage: The Influence of Gender Ideology on Housework Allocation Over the Life Course’, Journal of Family Issues 26(8): 1037–61.

18.

Davis

S.N.

Greenstein

T.N.

(2004) ‘Cross-national Variations in the Division of Household Labor’, Journal of Marriage and Family 66: 1260–71.

19.

Ferree

M.M.

(1991) ‘The Gender Division of Labor in Two-Earner Marriages: Dimensions of Variability and Change’, Journal of Family Issues 12(2): 158–80.

20.

Folbre

(2004) ‘A Theory of the Misallocation of Time’, pp. 7–24 in Folbre

Bittman

(eds) Family Time: The Social Organization of Care. London: Routledge.

21.

Gershuny

Godwin

Jones

(1994) ‘The Domestic Labour Revolution: A Process of Lagged Adaptation?’, pp.151–97 in Anderson

Bechofer

Gershuny

(eds) The Social and Political Economy of the Household. Oxford: Oxford University Press.

22.

Greenstein

T.N.

(2000) ‘Economic Dependence, Gender, and the Division of Labour in the Home: A Replication and Extension’, Journal of Marriage and the Family 62(2): 322–35.

23.

Gupta

(2006) ‘Her Money, Her Time: Women’s Earnings and Their Housework Hours’, Social Science Research 35: 975–99.

24.

Gupta

(2007) ‘Autonomy, Dependence or Display? The Relationship between Married Women’s Earnings and Housework’, Journal of Marriage and Family 69: 399–417.

25.

Hochschild

A.R.

(1989) The Second Shift. New York: Avon.

26.

Hook

J.L.

(2006) ‘Care in Context: Men’s Unpaid Work in 20 Countries, 1965–2003’, American Journal of Sociology 71(4): 639–60.

27.

Morehead

(2005) ‘Beyond Preference and Choice: How Many Mothers Allocate Time to Work and Family’, unpublished paper presented at the Families Matter AIFS Conference, Melbourne, 9–11 February.

28.

Office for Women (2006) Lifting and Raising the Bar. URL (consulted Hanuary 2009): www.aspc.gov.au/publications06/marriage.pdf

29.

Sayer

(2005) ‘Gender, Time and Inequality: Trends in Women’s and Men’s Paid Work, Unpaid Work and Free Time’, Social Forces 84(1): 285–303.

30.

Stata (2007) Stata Data Management Reference Manual Release 10. College Station, TX: Statacorp LP.

31.

Sullivan

(2000) ‘The Division of Domestic Labour: Twenty Years of Change?’, Sociology 34(3): 437–56.

32.

Sullivan

(2006) Changing Gender Relations, Changing Families: Tracing the Pace of Change Over Time. Lanham, MD: Rowman and Littlefield.

33.

West

Zimmerman

D.H.

(1987) ‘Doing Gender’, Gender & Society 1(2): 125–51.

34.

Western

J.S.

Emmison

J.M.

Baxter

Chant

D.C.

(1993) Social Structure of Australia Project, 1993. Data file held by Dept Anthropology and Sociology, University of Queensland, St Lucia.

35.

Western

J.S.

Boreham

P.R.

Clegg

Emmison

J.M.

Marks

G.N.

(2005) Social Structure of Australia Project, 1986. Australian Social Science Data Archive – UQ. Identification Number: au.edu.anu.assda.ddi.30001. URL (consulted February 2006): http://assda-nesstar.anu.edu.au/webview/index.jsp

36.

Western

Baxter

Pakulski

Tranter

Western

van Egmond

(2005) Neoliberalism, Inequality and Politics National Survey. Data file. St Lucia: The University of Queensland Social Research Centre.

37.

Whitehouse

(2004) ‘From Family Wage to Parental Leave: The Changing Relationship between Arbitration and the Family’, Journal of Industrial Relations 64(4): 400–412.